Zyte

What Zyte actually is

Zyte is the company behind Scrapy — the Python crawling framework it describes as "our open source web crawling framework written in Python", and which a very large share of production scrapers are still built on. That lineage matters, because it means Zyte is not a proxy reseller that bolted on an API. It has been solving the unglamorous parts of web data collection for years, and Zyte API is the commercial expression of that.

The product itself is deliberately narrow in surface and wide in coverage: one endpoint that handles ban avoidance, IP rotation across residential, datacenter and mobile pools, full headless-browser rendering, session persistence, geo-targeting, browser actions like clicks and form fills, and LLM-backed extraction that turns raw HTML into typed JSON. You do not choose a proxy type, buy rendering separately, or maintain your own retry ladder. You send a URL, and Zyte decides how much machinery that particular site needs.

The pricing model is the interesting part

Most scraping vendors sell you bandwidth and let you discover the hard way that a difficult site burns ten times more of it. Zyte prices by site difficulty instead, across five tiers, and charges only for what a request actually consumed.

On pay-as-you-go with no commitment, plain HTTP responses run $0.13 per 1,000 requests for Simple sites, rising through $0.23 (Easy), $0.44 (Moderate) and $0.70 (Complex) to $1.27 for Advanced. If a site needs a real browser, the same ladder runs $1.01 to $16.08 per 1,000. Committing to $100 a month cuts those to $0.10–$0.95 and $0.75–$12.00; at $200 a month they fall again to $0.08–$0.76 and from $0.60. Zyte's headline rate of $0.06 per 1,000 successful responses sits at the top of that commitment ladder.

Two things follow. First, your bill tracks the difficulty of your targets rather than the size of the pages, which is far easier to forecast once you know your target list. Second, the gap between the cheap and expensive ends is enormous — a 100x spread between Simple HTTP and Advanced browser rendering. If you are scraping a handful of stubborn, heavily-defended sites, model your costs at the top of the range before you commit, not the bottom.

Independent benchmarks back the performance claim

Vendor benchmarks are usually worthless. This one is not Zyte's own: Proxyway, an independent reviewer, tested the major unblocking APIs and found Zyte ahead on both axes at once.

• Zyte — 97.82% success rate, 11.70s average response
• Bright Data — 94.65%, 15.05s
• Smartproxy — 93.79%, 20.05s
• Oxylabs — 93.37%, 20.43s
• Crawlbase — 90.00%, 16.21s

Winning on success rate and latency simultaneously is unusual, and it compounds: a higher success rate means fewer paid retries, and since Zyte bills per successful response, the effective cost per usable record drops faster than the headline rate suggests. Three percentage points of success rate sounds trivial until you are running ten million requests a month.

Where it fits — and where it doesn't

Zyte suits teams who already know they have a scraping problem: data teams feeding AI training sets or RAG pipelines, e-commerce and pricing intelligence, job-board and real-estate aggregation, market research, and anyone whose existing Scrapy project keeps tripping over bans. The Scrapy Cloud product runs and monitors your spiders if you would rather not host them, and Zyte has begun shipping coding-agent add-ons that give tools like Claude Code the context to scaffold production Scrapy projects.

It is a poor fit if you want a no-code point-and-click scraper. There is no visual builder here — this is an API for engineers, and you will be writing code. If you need a handful of pages a week from simple sites, a $30/month no-code tool will serve you better. Zyte earns its place when volume, difficulty or reliability requirements make ad-hoc scraping untenable.

Trust, compliance and lock-in

For anyone collecting data commercially, the compliance story matters more than the feature list. Zyte is ISO 27001 certified and publishes the certificate, requires the same of its hosting providers, and commits to GDPR and CCPA. It also co-authored the FISD Alternative Data Standards — a rare signal that a scraping vendor is engaging with the regulated end of the market rather than avoiding it. Compliance-conscious buyers can hand that to legal without a long conversation.

Lock-in is unusually low for this category. The base tier has no minimum commitment, so you can stop whenever you like; the $100 and $200 monthly tiers are opt-in discounts, not contracts. More importantly, Scrapy is open source — your crawler logic stays yours, and moving your spiders elsewhere does not mean rewriting them from scratch.